Byron T. Ahn | Grammar of self-referring expressions

The distribution of reflexive pronouns in a language has a rich tradition of being explored in modern linguistic theory from the perspective of morphosyntax, semantics, and the syntax-semantics interface — especially for languages like English. This major branch of my research explores this well-studied domain with data from new perspectives: the syntax-prosody interface and the internal structure of English reflexive -self expressions. The aim of this research is to uncover new patterns that can inform the best way to model reflexive binding in a language like English.

A key conclusion reached throughout this domain of my (ongoing) research is that Binding Theory should not be conceptualized as confined to one linguistic module. Instead, what we call “binding” (the rules/constraints governing the distribution and form of reflexive pronouns) is instead the result of a variety of rules/constraints at different levels of the grammar (i.e., some syntactic components, plus some semantic components, plus some morphological components…): what I have (cheekily) called Distributed Binding Theory.

This has been informed by research that addresses the following questions:

Which -self anaphors are unstressed/stressed/focused in English? (prosody)
How do syntactic operations/representations influence anaphors’ prosody? (syntax-prosody interface)
What is the internal composition of X+SELF anaphors in English? (morphosyntax)
When can the phi-features of a X+SELF expression differ from those of the antecedent? (morphosyntax)
What determines the acceptability of themself vs. themselves, with [sg] antecedents? (experimental sociosyntax)

These research questions are taken up in my own work, as well as collaborative work with Laura Kalin and Kirby Conrod.

In the future, this program aims to build a better understanding of the grammar of phi-matching in English (arguing against models that require Agree between antecedent of binding and anaphor) as well as of the syntactic components of binding (exploring how the internal functional structure of an anaphoric nominal may impact its clausal distribution).

some work in this project

Conrod, Kirby, Ruth Schultz & Byron Ahn. In Press. How many selves for them? In NELS 52: Proceedings of the 52nd Annual Meeting of the North East Linguistics Society,
abstract Paper Materials Bib

This work investigates the syntactic and semantic distribution of reflexive forms themself vs. themselves. We focus on whether the use of -self vs -selves is affected by specificity of the antecedent, distinguishing seven levels for a singular antecedent: quantified indefinites (some/any), quantified universals (each/every/no), generic kind definites, specific indefinites, specific definites, unknown specific person, named person. We administered an online acceptability judgment survey to 1,127 participants, recruited on Twitter and Prolific, who rated sentences on a 5-point Likert scale from “very natural” to “very unnatural,” in a randomized order, and then completed a short demographic survey. We find that antecedent type, -self/-selves form of the anaphor, and social variables including age, gender, and endorsement of prescriptive language ideology all affected how participants rated the sentences; themself was generally rated higher than themselves for most antecedents, contra previous findings. Our results seem to indicate that themself is preferred by many English speakers for singular antecedents, and also that the singular use of themself/themselves is a sociosyntactically active and salient variable across those speakers.
@incollection{Conrod-Schultz-Ahn-NELS52, author = {Conrod, Kirby and Schultz, Ruth and Ahn, Byron}, booktitle = {{NELS} 52: Proceedings of the 52nd Annual Meeting of the {N}orth {E}ast {L}inguistics {S}ociety}, title = {How many selves for them?}, year = {In Press} }
Conrod, Kirby & Byron Ahn. 2023, January. Who accepts themself? Sociosyntactic variation in English -self reflexives. Talk presented at the 2023 annual meeting of the LSA. Denver, CO.
abstract Bib

Though sg. they has been recently undergoing a change with respect to the types of antecedents it can take, it has been well attested throughout the history of English (Bjorkman 2017, Conrod 2019, Konnelly and Cowper 2020, Whitley 1978). Also well attested throughout the history of English, there are two reflexive forms for sg. they: themself and themselves (as noted in Whitley 1978). The present study explores what factors (social, ideological, and linguistic) influence acceptability of these forms. While preceding work has found themselves to be more acceptable in sg. they contexts (Ackerman et al. 2018), the present study uncovers more nuanced patterns. We find three patterns of acceptability, which we infer to represent different dialects. The parameters underlying these dialects have to do with how feature matching in binding can vary across English grammars.
@misc{conrodahn23, author = {Conrod, Kirby and Ahn, Byron}, howpublished = {Talk presented at the 2023 annual meeting of the LSA. Denver, CO}, title = {Who accepts \emph{themself}? Sociosyntactic variation in English \emph{-self} reflexives}, year = {2023} }
Ahn, Byron & Kirby Conrod. 2022, March. Three ways to rate themself. Talk presented at 35th Annual Conference on Human Sentence Processing (formerly the ‘CUNY Conference on Human Sentence Processing’). University of California, Santa Cruz.
abstract Slides Bib

Speakers’ acceptance of they referring to a singular antecedent (“singular they”) correlates with linguistic and social factors (e.g., antecedent type, or speaker age, gender, and ideology). The present study (i) investigates how themselves/themself are rated with different syntactically singular antecedent types and (ii) identifies three distinct dialect groups based solely on those ratings.

Importantly, the variation in ratings is not only impacted by factors like social variables and linguistic ideologies, but it is more greatly impacted by dialect groups that do not reduce to these factors. These results help document language variation/change, inform understanding of the variable grammar(s) of singular they, and furnish further evidence that the phi-features in English reflexive nominals do not need to fully match those of their anteceden.
@unpublished{Ahn-Conrod-2022-themself, author = {Ahn, Byron and Conrod, Kirby}, howpublished = {Talk presented at 35th Annual Conference on Human Sentence Processing (formerly the ‘CUNY Conference on Human Sentence Processing’). University of California, Santa Cruz}, title = {Three ways to rate themself}, year = {2022} }
Ahn, Byron. 2019, April. Anaphor Binding and Phi-(Mis)Matches in English. Talk presented at DISCO (Dependency in Syntactic Covariance): Phi-agreement, reference and case. Universiy of Leipzig.
abstract Handout Bib

English reflexive anaphors, such as ourselves, are derived syntactic objects, with internal structure paralleling possessives (and are not statically listed as morphosyntactically simplex [+reflexive] objects in the lexicon; Ahn and Kalin 2018, also Postal 1966). The pronominal component of the English anaphor is widely described (cf. discussion in Sundaresan 2018) as needing to φ-match the antecedent of binding. In other words, the generalization in (1) is seen as describing the fact that ourselves must be used in (2), and not themselves. (Features of the binder are diagnosed by, e.g., nominal and verbal morphology.)

(1) Generalization about φ-matching in English reflexive anaphors
Pronouns in a reflexive anaphor φ-match the local antecedent of binding

(2) We will not embarrass ourselves / *themselves

While (1) is generally true, φ-mismatch is possible between the binder and bound pronoun: two contexts where mismatch is possible are given in (3).

(3) a. The U.N. finds themselves / itself in a difficult position
b. Does Your Honor doubt yourself / herself?

At the same time, there are cases where mismatch is ruled out; two such cases are in (4), which are minimal pairs with those in (3).

(4) a. The U.N. countries find themselves / *itself in a difficult position
b. Do you doubt yourself / *herself?

This talk explores two questions. First, empirically, in what contexts can φ-mismatches (not) occur? Using data from informal judgments, corpus searches, and a new experiment, we will find that φ-mismatches are constrained in generalizable ways. Second, theoretically, what types of theory reflexive binding are plausible, in light of these generalizations? Some have argued that English reflexive anaphors are φ-deficient, requiring φ-valuation by the local antecedent of binding (perhaps mediated by a functional head, Krazter 2009); however, these findings suggest that English pronouns in reflexives exhibit a split in this area, between 3.sg and others. I suggest this is related to 3.sg pronouns being the only ones in English that mark gender.
@unpublished{Ahn-DISCO, author = {Ahn, Byron}, howpublished = {Talk presented at DISCO (Dependency in Syntactic Covariance): Phi-agreement, reference and case. Universiy of Leipzig}, title = {Anaphor Binding and Phi-(Mis)Matches in {E}nglish}, year = {2019} }
Ahn, Byron. 2019. Features, Identity, and ‘Yourself.’ In Maggie Baird & Jonathan Pesetsky (eds.), NELS 49: Proceedings of the 49th Annual Meeting of the North East Linguistics Society, Vol. 1, 15–24.
abstract Paper Handout Bib

It has been widely assumed that the φ-features of bound pronouns must φ-match their antecedents. (We behave ourselves/*themselves.) While generally true, φ-mismatches are regularly (and commonly) attested, e.g.,:

Each of us behaves ourselves/themselves. (3sg antecedent; 1pl/3pl anaphor)

I wouldn’t blame yourself.1em (1sg antecedent; 2sg anaphor)

However, mismatch is constrained in informative ways. The patterns indicate that there are two derivational pathways for bound pronouns, such that 3.sg bound pronouns necessarily involve extra derivational machinery.
@incollection{Ahn-NELS49, author = {Ahn, Byron}, booktitle = {{NELS} 49: Proceedings of the 49th Annual Meeting of the {N}orth {E}ast {L}inguistics {S}ociety}, editor = {Baird, Maggie and Pesetsky, Jonathan}, pages = {15--24}, title = {Features, Identity, and `Yourself'}, volume = {1}, year = {2019} }
Ahn, Byron & Laura Kalin. 2018. What’s in a (English) Reflexive? In NELS 48: Proceedings of the 48th Annual Meeting of the North East Linguistics Society, Vol. 1, 1–13.
abstract Paper Handout Bib

In this paper, we investigate the internal structure of English reflexive anaphors. While it is clear that English Reflexive anaphors consist of a pronoun and a form of “self”, any questions of how their internal structure is derived have been largely ignored. We take up two puzzles the explore this more deeply.

First, the case puzzle: the pronominal components are genitive in 1st/2nd person (myself, yourselves) but accusative in 3rd person (himself, themselves).

Second, the modifier puzzle: modifiers are possible between the pronoun and “self”, but only some of them still permit the expression to occur in reflexive positions (Behave your annoying/*tired self!, She her brilliant/*caffeinated selfsolved a Millennium Problem.).

We show that the solutions to these puzzles are inter-related, and collude to lead us towards a precise syntactic analysis. In particular, our investigations into these puzzles have led us to make several conclusions: (i) they are syntactically complex, (ii) their reflexivity emerges through the derivation, and (iii) at least some morphological processes must ordered after syntax. Moreover, this work may have serious implications for many other areas, such as the analysis of exempt reflexives, what varies across dialects, and what varies across languages.
@incollection{Ahn-Kalin-NELS48, author = {Ahn, Byron and Kalin, Laura}, booktitle = {{NELS} 48: Proceedings of the 48th Annual Meeting of the {N}orth {E}ast {L}inguistics {S}ociety}, pages = {1--13}, title = {What's in a ({E}nglish) Reflexive?}, volume = {1}, year = {2018} }
Ahn, Byron. 2017, May. (In Search of) Universals in Reflexive Syntax. Talk presented at Cambridge Comparative Syntax 6.
abstract Handout Bib

Given that “reflexivity” is known to be a heterogeneous class (local vs. long-distance; exempt vs. bound; subject-oriented vs. non-subject oriented; etc.), what are the prospects for true universals in reflexivity? In this talk, I focus on Local Subject-Oriented Reflexivity (LSOR), its properties, and how to derive them.

By studying the formal constraints on LSOR markers, as well as their varied crosslinguistic realization, I argue that: (i) for an LSOR derivation that involves a reflexive functional head on the clausal spine, (ii) an argument-saturating nominal moves to the specifier of that projection in LSOR derivations, and (iii) languages can vary in which of these components are observable.

Given the apparently wide cross-lingusitic applicability of this derivation, this suggests that binding in LSOR is not simply a special type of normal binding conditions, but a separate phenomenon which is available across languages, due to what is made available by a shared universal linguistic competence.

The question remains, then, what is the nature of precisely what it is that is made available by UG, and why children acquiring language in so many contexts readily acquire the same (or very similar) LSOR structures. In the context of the ReCoS research group, I engage with the idea of emergent parameters and suggest that we need to reconsider “reflexive” patterns as the result of a single learning path.
@unpublished{Ahn-CamCoS6, author = {Ahn, Byron}, howpublished = {Talk presented at Cambridge Comparative Syntax 6}, title = {(In Search of) Universals in Reflexive Syntax}, year = {2017} }
Ahn, Byron. 2016. Reflexes of Reflexivity: Locality and the Interfaces. Proceedings from the 50th Annual Meeting of the Chicago Linguistic Society.
abstract Paper Bib

When a reflexive argument is bound by a local subject (Local Subject-Oriented Reflexivity, LSOR), many languages employ a distinct morphosyntactic configuration for marking reflexivity, including: Dogrib (Na-Dené) ede, Italian (Romance) si, Japanese (Altaic) zibunzisin, Kannada (Dravidian) -koND, Russian (Slavic) sebe, and Shona (Niger-Congo) zvi.

This raises an important and naive question: what is it about local subjects that gives them a privileged ability to license LSOR-marking? Upon further investigation, a second more informed question arises: why do only some local subjects (specifically, non-derived subjects) license LSOR-marking?

I argue that reflexive anaphors move to a position near the subject, and this movement is done to locally satisfy the syntactic and semantic demands of a reflexive Voice⁰, REFL, which ensures (without stipulation) that the subject is not a derived one. With only REFL, its syntactic/semantic properties, and constraints on locality, we thus derive how the grammatical role of the antecedent may effect LSOR-marking.

Additionally, this analysis derives the range of possibilities for morphosyntactic exponents of LSOR: including a unique anaphor, verbal affix, and/or special word-order. Specifically, REFL Voice⁰ can be realized (or not), it induces a (covert) movement that can result in a special word-order (or not), and anything it selects (including an anaphor) can be featurally-distinct and can thus have a unique form (or not). These are all purely instances of surface variation with the same formal underpinnings.

As such, the subject-orientation puzzles presented by LSOR, as well as its possible morphosyntactic instantiations, are reduced to issues of locality and surface variation.
@article{Ahn-CLS50, author = {Ahn, Byron}, journal = {Proceedings from the 50th Annual Meeting of the {C}hicago {L}inguistic {S}ociety}, title = {Reflexes of Reflexivity: Locality and the Interfaces}, year = {2016} }
Ahn, Byron. 2015. Giving Reflexivity a Voice: Twin Reflexives in English. UCLA dissertation.
abstract Thesis Bib

Across languages, there is variability in the surface realization of reflexivity, according to various sets of properties. For example, there are languages (e.g. Greek, Lakhota) that seem to treat some of their reflexive clauses as being in a non-active voice, similar to a passive. There are also languages (e.g. French, Kannada) in which reflexivity is encoded differently depending on whether the antecedent is the subject or not. In this way, English seems to be different: reflexivity is apparently realized in a homogeneous way – filling an argument position with an anaphoric expression like themselves – regardless of clausal voice or grammatical role of the antecedent.

This homogeneity is an illusion. Despite using a single set of anaphoric expressions for reflexivity in various situations, reflexive anaphors in English fall into two classes: those that exhibit exceptional prosodic behaviors, and those that do not.

This exceptionality can be directly observed in two domains: the distribution of "default" phrasal stress, and the distribution of a certain focal accent. From the results of investigations on these issues, I show that the distribution of exceptionally behaving reflexive anaphors is structurally constrained. This implicates that there must a be syntactic account for these prosodic properties.

Assuming that syntactic structure plays a near deterministic role in prosody (an assumption going back to even the earliest generative work on phrasal stress; Chomsky and Halle 1968:25), I argue for a more refined syntactic structure of reflexivity. Briefly, I demonstrate a sub-class of reflexive anaphors in English undergo a syntactic movement (to a reflexive VoiceP). This movement, along with independently motivated mechanisms for placement of phrasal stress and focal accents, derives the heterogeneous prosodic behaviors of reflexives in English. Crucially, this analysis does not require the prosodic component to have any stipulations for specific (classes of) words, in line with a Minimalist approach to the Syntax-Prosody Interface.

This model of reflexivity simultaneously reduces the amount of theoretical machinery necessary to achieve descriptive adequacy, while also enhancing the model’s predictive power. Moreover, this research has broad theoretical implications, beyond just reflexives in English. This theory is able to unify the various morpho-syntactic instantiations of reflexivizing functions – across languages – as being related to the Reflexive VoiceP. It also establishes a core set of properties that define clausal reflexivity, each of which are the result of the formal properties of the reflexive Voice head. Finally, it provides direct support for the hypothesis that syntactic and prosodic structures are maximally isomorphic, with prosodic cues in the signal giving direct evidence for otherwise invisible in syntactic structure.
@phdthesis{Ahn-PhDThesis, author = {Ahn, Byron}, school = {UCLA}, thesis = {https://ling.auf.net/lingbuzz/002359}, title = {Giving Reflexivity a Voice: Twin Reflexives in {E}nglish}, year = {2015} }
Ahn, Byron & Dominique Sportiche. 2014. Bind Locally Indeed. In Carson T. Schütze & Linnaea Stockall (eds.), Connectedness: Papers by and for Sarah VanWagenen, Vol. 18, 61–69.
abstract Paper Bib

We present new evidence supporting the generalization encoded in rule H, a rule defended in Fox (2000): this rule encodes an economy principle on pronominal binding according to which a closer binder bleeds binding by a more distant one if both would yield the same semantic representa- tion. The evidence we discuss is based on how English reflexive pronouns behave in two different cases: one involving focus and one involving the (non)availability of phrasal stress.
@incollection{Ahn-Sportiche-BindLocally, author = {Ahn, Byron and Sportiche, Dominique}, booktitle = {Connectedness: Papers by and for {S}arah {V}an{W}agenen}, editor = {Sch\"utze, Carson T. and Stockall, Linnaea}, pages = {61--69}, series = {UCLA Working Papers in Linguistics}, title = {Bind Locally Indeed}, volume = {18}, year = {2014} }
Ahn, Byron. 2014, April. Reflexes of Reflexivity: Locality and the Interfaces. Talk presented at CLS 50, University of Chicago.
abstract Handout Bib

When a reflexive argument is bound by a local subject (Local Subject-Oriented Reflexivity, LSOR), many languages employ a distinct morphosyntactic configuration for marking reflexivity, including: Dogrib (Na-Dené) ede, Italian (Romance) si, Japanese (Altaic) zibunzisin, Kannada (Dravidian) -koND, Russian (Slavic) sebe, and Shona (Niger-Congo) zvi.

This raises an important and naive question: what is it about local subjects that gives them a privileged ability to license LSOR-marking? Upon further investigation, a second more informed puzzle arises: only some local subjects (specifically, non-derived subjects) license LSOR-marking.

I argue that reflexive anaphors move to a position near the subject, and this movement is done to locally satisfy the syntactic and semantic demands of a reflexive Voice⁰, REFL, which ensures (without stipulation) that the subject is not a derived one. With only REFL, its syntactic/semantic properties, and constraints on locality, we thus derive how the grammatical role of the antecedent may effect LSOR-marking.

Additionally, this analysis derives the range of possibilities for morphosyntactic exponents of LSOR: including a unique anaphor, verbal affix, and/or special word-order. Specifically, REFL Voice⁰ can be realized (or not), it induces a (covert) movement that can result in a special word-order (or not), and anything it selects (including an anaphor) can be featurally-distinct and can thus have a unique form (or not). These are all purely instances of surface variation with the same formal underpinnings.

As such, the subject-orientation puzzles presented by LSOR, as well as its possible morphosyntactic instantiations, are reduced to issues of locality and surface variation.
@unpublished{Ahn-cls50-refl, author = {Ahn, Byron}, howpublished = {Talk presented at CLS 50, University of Chicago}, title = {Reflexes of Reflexivity: Locality and the Interfaces}, year = {2014} }
Ahn, Byron. 2013, July. Universality and Subject-Oriented Reflexivity. Talk presented at the ICL 19, Geneva, Switzerland.
abstract Handout Bib

Reflexive anaphors are not a homogeneous class – there are local, long-distance, exempt, subject-oriented, non-subject oriented (perhaps among others) – and each different subtype exhibits different properties. I focus on Local Subject-Oriented Reflexivity (LSOR), its properties, and how to derive them. In approaching LSOR, I pose two puzzles.

First, a naive one: why do subjects license LSOR? Second, a more informed one: why only some subjects (i.e. non-derived subjects) possible licensers of LSOR?

I propose a novel solution to these puzzles which has two main components: (i) a semantic reflexivizer (associated with a unique grammatical Voice head, REFL), and syntactic movement of the anaphor (triggered by that same semantic reflexivizer in REFL).

In addition to solving these puzzles, this analysis correctly predicts the variation one finds in LSOR marking across languages. It also suggests that binding in LSOR is not simply a special type of normal binding conditions, but a separate phenomenon which is universally grammatically active, due to what UG makes available to all languages.
@unpublished{Ahn-icl19, author = {Ahn, Byron}, howpublished = {Talk presented at the ICL 19, Geneva, Switzerland}, title = {Universality and Subject-Oriented Reflexivity}, year = {2013} }
Ahn, Byron. 2012, January. The Prosody of Binding: Reflexive Voice and Default Sentential Stress. Talk presented at the 2012 annual meeting of the LSA, Portland, OR.
abstract Handout Bib

Syntactic structure has long been recognized as playing a crucial role in calculating Default Sentential Stress (DSS). Reflexive pronouns introduce an apparent problem: they tend not to bear DSS where other DPs would ("Remy SAW himself"/"Remy saw MARIE"). However, reflexives do not always avoid DSS; manipulating syntactic context, a reflexive can bear DSS (e.g., "Remy saw Marie and HIMSELF"). Based on such data, I argue against a string-positional account of DSS (e.g. SPE’s Nuclear Stress Rule). Moreover, I present evidence for the existence of subject-oriented reflexives in English, and motivate substantial revisions to Binding Theory.
@unpublished{Ahn-lsa2012, author = {Ahn, Byron}, howpublished = {Talk presented at the 2012 annual meeting of the LSA, Portland, OR}, title = {The Prosody of Binding: Reflexive Voice and Default Sentential Stress}, year = {2012} }
Ahn, Byron. 2012. External Argument Focus and the Syntax of Reflexivity. Coyote Papers: Working Papers in Linguistics, Linguistic Theory at the University of Arizona 20.
abstract Paper Bib

A sentence like Johnny burned himself, with the sole focal stress on the reflexive anaphor, can be felicitously interpreted as a response to a subject-question like Who burned Johnny?. This is unexpectedly felicitous, given previous accounts of focus in question-answer pairs. Critically, this phenomenon is limited to a subclass of reflexive anaphors. Though commonly accepted theories of focus and reflexivity would not predict this to be possible, by amending the syntactic representation of reflexivity, this phenomenon and its particular constraints are entirely predictable on the basis of normal rules of syntax-prosody mapping. This analysis gains independent support from the distribution of phrasal stress with regard to reflexive anaphors, as well as from the realization of reflexivity of other languages.
@article{Ahn-WCCFL29, author = {Ahn, Byron}, journal = {Coyote Papers: Working Papers in Linguistics, Linguistic Theory at the University of Arizona}, title = {External Argument Focus and the Syntax of Reflexivity}, volume = {20}, year = {2012} }
Ahn, Byron. 2010, September. Syntactic Configurations of Emphatic Reflexives. Talk presented at the workshop on Peculiar Binding Configurations, University of Stuttgart.
abstract Handout Bib
Many languages seem to allow reflexive pronouns to function either as arguments (1) or as adjuncts (2). The adjunct-type is called an Emphatic Reflexive (ER).

(1) a. Kenneth injured himself.
b. They bought themselves some time.
(2) a. No boy ate all the Cheezy Blasters himself.
b. Liz herself grew up in White Haven.

There are exactly two lexically differentiated ERs:
- ^vpER, loosely paraphrasable as ‘without help’ (2a)
- ^dpER, loosely paraphrasable as ‘X, not Y’ (2b)
Each has its own syntactic configuration, as well as its own felicity conditions that sensitive to:
- The type of the associate DP, or
- the argument structure of the predicate.
The ER reflexive pronoun is bound within a very local domain, even when the associate DP appears rather distant/disjoint
- ^dpERs adjoin to the DP and may be stranded by A-movement, as was argued for Floated Quantifiers in Sportiche (1988)
- ^vpERs are right-adjoined in the near-immediate c-command domain of the associate DP’s merge position
Languages in which ERs do not involve reflexive pronouns (e.g. German) are only minimally different from languages which do (e.g. English)
- Reflexive pronouns represent a syntactically obligatory, semantically null argument (similar to cases like perjure oneself) of an ER⁰
- a. German-type: [erP [er⁰ selbst ]]
  b. English-type: [erP [er⁰ Ø ] [DP himself ]]
@unpublished{Ahn-pbc-er, author = {Ahn, Byron}, howpublished = {Talk presented at the workshop on Peculiar Binding Configurations, University of Stuttgart}, title = {Syntactic Configurations of Emphatic Reflexives}, year = {2010} }
Ahn, Byron. 2010. Not Just Emphatic Reflexives Themselves: Their Prosody, Semantics and Syntax. UCLA dissertation.
abstract Thesis Bib

Emphatic Reflexives is the term used to describe the use of reflexive pronouns in an emphatic, adjunct-like way, as in "Sayoko drove to LA herself." Emphatic Reflexives (ERs) have been investigated by many in the course of generative linguistics (Moravcsik 1972, Edmondson and Plank 1978, Browning 1993, Eckardt 2001, Hole 2002, Konig and Gast 2002, Bergeton 2004, Konig and Siemund 2005, and Tavano 2006, to name a few), but past analyses leave some questions inadequately or not at all answered.

I endeavor to explain as many of the phenomena as possible, focusing on three main points. First, there are exactly two readings of ERs, each with its own felicity conditions, syntactic licensing and semantic properties. Second, ERs are indeed instances of reflexive pronouns that are subject to the Binding Conditions; and though some ERs may appear linearly disjoint from their associate DP, they are always (extremely locally) bound. Thirdly, all ERs are obligatorily contrastively focused under both readings, and they exhibit prosodic properties related to this.

In making these arguments, I also find evidence for hypotheses that may have broader impact, such as the fact that the Voice head must be the licenser of volitional agents and that the MAE ToBI model of English intonational structure needs modification.
@mastersthesis{Ahn-MAThesis, author = {Ahn, Byron}, school = {UCLA}, thesis = {Ahn-MA.pdf}, title = {Not Just Emphatic Reflexives Themselves: Their Prosody, Semantics and Syntax}, year = {2010} }

	Each of us behaves ourselves/themselves.	(3sg antecedent; 1pl/3pl anaphor)
	I wouldn’t blame yourself.1em	(1sg antecedent; 2sg anaphor)